Advertisement

Quantitative and qualitative detailed milk protein profiles of 6 cattle breeds: Sources of variation and contribution of protein genetic variants

Open ArchivePublished:October 14, 2020DOI:https://doi.org/10.3168/jds.2020-18497

      ABSTRACT

      Different fractions of milk nitrogenous compounds (not only caseins) have different effects on the nutritional value of milk, its coagulation and curd firming properties, and its cheese-making efficiency. To assess different sources of variation, especially the cows' breed and genetic variants of the main protein fractions, milk samples were collected from 1,504 cows belonging to 3 dairy breeds (Holstein-Friesian, Brown Swiss, and Jersey) and 3 dual-purpose breeds (Simmental, Rendena, and Alpine Grey) reared in 41 multibreed herds. Beyond crude protein, casein (CN), and urea, 7 protein fractions were analyzed using HPLC, and 5 other N fraction traits were calculated. All 15 traits were measured qualitatively (% of milk N) and quantitatively (g/L of milk). The HPLC technique allowed us to discriminate between the main genetic variants of β-CN, κ-CN, and β-lactoglobulin and thus to genotype the cows for the CSN2, CSN3, and BLG genes, respectively. Data were analyzed using 2 mixed models, both including the effects of herd-date, breed, parity, and lactation stage, and only one also including the effects of the genotypes of the milk proteins. Breed of cow explained 2 to 36% of phenotypic variability for all the N fractions, with the exception of the urea and total casein contents of milk and the urea and β-CN proportions of total milk N. Lactation stage had a considerable influence on the amount (g/L) of almost all the protein fractions in milk, but neither the nonprotein N fractions nor the percentage of milk N protein profile were affected. The inclusion of the CSN2, CSN3, and BLG genotypes in the model explained a large part of the total variability in all the milk protein and nonprotein fractions except urea. It also reduced the variance explained by breed and residual factors. An exception was shown by the proportion of αS1-CN variance explained by breed that moved from 13 to 28%. Similarly, for amount (g/L) of β-CN, the effect of breed became significant (12%), whereas it was almost null before inclusion of genotypes. In terms of percentage of milk N, the genotypes of CSN3 notably affected all the casein fractions, whereas the BLG genotypes had a much greater influence on most noncasein traits. The genotypes of the CSN2 gene exerted an appreciable effect on αS2-CN and not β-CN, as expected. Comparing the 2 models, we were also able to discriminate the effect of the breed on a milk N fraction, both quantitatively and qualitatively, in 2 quotas: the first due to the milk protein polymorphisms (major genes) and the second due to other genetic factors (polygene), after correcting for the effect of herd-date of sampling, parity, and lactation stage. The knowledge about the detailed milk protein profile of different cattle breeds provided by this study could be of great benefit for the dairy industry, providing new tools for the enhancement of milk payment systems and breeding program designs.

      Key words

      INTRODUCTION

      Milk protein composition is an important factor determining the nutritional and technological properties of milk. Bovine milk is characterized by having proteins with a high biological value, which are sources of essential amino acids and bioactive peptides. These molecules provide the organism with nutrients and benefit the body's physiological functions with their antimicrobial, antioxidant, antithrombotic, antihypertensive, and immunomodulatory activities (
      • Korhonen H.
      • Pihlanto A.
      Bioactive peptides: Production and functionality.
      ). The protein profile of bovine milk is very different from that of human milk, particularly for the lower amount of CP and especially of caseins (
      • Barłowska J.
      • Szwajkowska M.
      • Litwińczuk Z.
      • Kròl J.
      Nutritional value and technological suitability of milk from various animal species used for dairy production.
      ). In both species, every nitrogenous compound has its specific biological function (
      • Haug A.
      • Høstmark A.T.
      • Harstad O.M.
      Bovine milk in human nutrition—A review.
      ).
      With regard to the technological properties of milk, casein fractions (the main ones being αS1-, αS2-, β-, and κ-CN) are directly involved in the cheese-making process, given that they are the major components of the curd, together with milk fat. The whey protein fractions (principally β-LG and α-LA) are not trapped in the curd and are lost in the whey. However, they can be recovered in ricotta-type cheeses or in other heat-acid cheese-making processes (
      • Law B.A.
      • Tamime A.Y.
      Technology of Cheese-Making.
      ). Many other minor protein fractions are known to be present in milk, such as di- and polypeptides, free amino acids, and NPN compounds (
      • Wittenburg D.
      • Melzer N.
      • Willmitzer L.
      • Lisec J.
      • Kesting U.
      • Reinsch N.
      • Repsilber D.
      Milk metabolites and their genetic variability.
      ).
      Several factors influence the protein profile of bovine milk, such as season, parity, stage of lactation and physiological status of the cows, and their feeding regimen (
      • Bobe G.
      • Beitz D.C.
      • Freeman A.E.
      • Lindberg G.L.
      Effect of milk protein genotypes on milk protein composition and its genetic parameter estimates.
      ), although genetic factors play the most important role (
      • Ng-Kwai-Hang K.F.
      • Hayes J.F.
      • Moxley J.E.
      • Monardes H.G.
      Variation in milk protein concentrations associated with genetic polymorphism and environmental factors.
      ;
      • Heck J.M.L.
      • Olieman C.
      • Schennink A.
      • van Valenberg H.J.F.
      • Visker M.H.P.W.
      • Meuldijk R.C.R.
      • van Hooijdonk A.C.M.
      Estimation of variation in concentration, phosphorylation and genetic polymorphism of milk proteins using capillary zone electrophoresis.
      ). As is well known, milk obtained from different breeds is characterized by different contents of the major protein fractions (
      • Poulsen N.A.
      • Jensen H.B.
      • Larsen L.B.
      Factors influencing degree of glycosylation and phosphorylation of caseins in individual cow milk samples.
      ). Within the same breed, different cows may produce milk with different genetic variants of the main protein fractions, and in different breeds the frequencies of these genetic variants are not the same (
      • Caroli A.M.
      • Chessa S.
      • Erhardt G.J.
      Invited review: Milk protein polymorphisms in cattle: Effect on animal breeding and human nutrition.
      ). Moreover, the concentration of a given protein fraction in milk and its genetic variants are not independent of each other, as different genetic variants may have different levels of penetrance, causing different concentrations of that protein fraction in the milk (
      • Heck J.M.L.
      • Schennink A.
      • Van Valenberg H.J.F.
      • Bovenhuis H.
      • Visker M.H.P.W.
      • Van Arendonk J.A.M.
      • Van Hooijdonk A.C.M.
      Effects of milk protein variants on the protein composition of bovine milk.
      ). In fact, polymorphism in the genes coding for the major milk protein fractions is responsible for qualitative and quantitative changes in the milk protein profile. For example, the B variant of κ-CN increases the expression of the κ-CN gene (CSN3), resulting in a higher amount and proportion of κ-CN than with the A variant, and in the case of BLG the A variant is associated with a higher content of β-LG compared with the B variant (
      • Robitaille G.
      • Britten M.
      • Morisset J.
      • Petitclerc D.
      Quantitative analysis of β-lactoglobulin A and B genetic variants in milk of cows β-lactoglobulin AB throughout lactation.
      ;
      • Hallén E.
      • Wedholm A.
      • Andrén A.
      • Lundén A.
      Effect of β-casein, κ-casein and β-lactoglobulin genotypes on concentration of milk protein variants.
      ). This means that it is not well known whether the different properties observed for milk containing A-κ-CN or B-κ-CN (or containing A-β-LG or B-β-LG) depend on physicochemical differences between the 2 κ-CN (β-LG) variants or depend on the effect of the 2 alleles on the quantity and proportion of κ-CN (β-LG) in milk. Last, as reviewed by
      • Bittante G.
      • Penasa M.
      • Cecchinato A.
      Invited review: Genetics and modeling of milk coagulation properties.
      , both the quantities and the genetic variants of the major protein fractions affect the technological properties of milk. Yet the effect of the protein fraction concentrations and of the major genes' genotypes have seldom been studied together to distinguish their respective effects and interactions. Recently, we quantified the effects of the concentrations of the main protein fractions, independently of the effects of their major genetic variants, on milk coagulation, curd firming, and syneresis (
      • Amalfitano N.
      • Cipolat-Gotet C.
      • Cecchinato A.
      • Malacarne M.
      • Summer A.
      • Bittante G.
      Milk protein fractions strongly affect the patterns of coagulation, curd firming, and syneresis.
      ) and on cheese yield and the recovery of milk nutrients in the cheese (
      • Cipolat-Gotet C.
      • Cecchinato A.
      • Malacarne M.
      • Bittante G.
      • Summer A.
      Variations in milk protein fractions affect the efficiency of the cheese-making process.
      ) of milk from Brown Swiss cows.
      Different breeds of cows are known to produce milk with different coagulation, curd firming, and syneresis properties (
      • Stocco G.
      • Cipolat-Gotet C.
      • Bobbo T.
      • Cecchinato A.
      • Bittante G.
      Breed of cow and herd productivity affect milk composition and modeling of coagulation, curd firming, and syneresis.
      ) and with different cheese yields, nutrient recoveries in the cheese, and cheese-making efficiencies (
      • Stocco G.
      • Cipolat-Gotet C.
      • Gasparotto V.
      • Cecchinato A.
      • Bittante G.
      Breed of cow and herd productivity affect milk nutrient recovery in curd, and cheese yield, efficiency and daily production.
      ). Different breeds of cow are also known to differ in the frequencies of the genetic variants of the major milk protein fractions (
      • Gustavsson F.
      • Buitenhuis A.J.
      • Johansson M.
      • Bertelsen H.P.
      • Glantz M.
      • Poulsen N.A.
      • Lindmark Månsson H.
      • Stålhammar H.
      • Larsen L.B.
      • Bendixen C.
      • Paulsson M.
      • Andrén A.
      Effects of breed and casein genetic variants on protein profile in milk from Swedish red, Danish Holstein, and Danish Jersey cows.
      ) and in the contents of these protein fractions (
      • Poulsen N.A.
      • Jensen H.B.
      • Larsen L.B.
      Factors influencing degree of glycosylation and phosphorylation of caseins in individual cow milk samples.
      ). However, the different breeds studied so far have often been reared on different farms, making it difficult to disentangle the effects of breed from those of rearing conditions.
      To the best of our knowledge, very few studies to date have been carried out characterizing the variability in the detailed milk protein profiles (obtained through gold standard methods) of cows of different breeds reared in the same multibreed herds. Moreover, even fewer studies have attempted to distinguish between the differences among breeds due to the frequencies of the genetic variants and the differences due to other (genetic) factors, while at the same time taking into account the effects of the major phenotypic factors affecting the protein fractions content of milk (e.g., herd, parity, lactation stage).
      For the present work we built detailed protein profiles of bovine milk (15 different traits) in which the various compounds are expressed in terms of content per unit of milk volume (g/L) and as proportions of total milk N (%N). We were able to assess the genotypes of the major genetic variants of some of the protein fractions (β-CN, κ-CN, β-LG). Therefore, the main aims of our study were (1) to assess the effect of the main sources of variation (herd, breed, parity, lactation stage, residual) for every milk protein trait, investigating also how the genotypes of the 3 major genes influence the main sources of variation, particularly breed, for each milk protein trait, and (2) to compare the protein profiles of milk from 3 dairy breeds (Holstein-Friesian, Brown Swiss, Jersey) and 3 dual-purpose breeds (Simmental, Rendena, Alpine Grey) reared in multibreed herds, both with and without inclusion in the model of the effects of the genotypes of the major genetic variants of some of the protein fractions.

      MATERIALS AND METHODS

      Animals and Sample Collection

      The present work is part of the Cowplus project. Herds were selected in the provinces of Trento and Bolzano (Bozen-Süd Tirol) in northeast Italy, giving preference to those rearing cows of different breeds (2–5 breeds per herd, on average 3.0 ± 0.8). The detailed characteristics of the dairy systems and the individual farms in the area were reported in previous studies within the same project (
      • Bobbo T.
      • Ruegg P.L.
      • Stocco G.
      • Fiore E.
      • Gianesella M.
      • Morgante M.
      • Pasotto D.
      • Bittante G.
      • Cecchinato A.
      Associations between pathogen-specific cases of subclinical mastitis and milk yield, quality, protein composition, and cheese-making traits in dairy cows.
      ;
      • Stocco G.
      • Cipolat-Gotet C.
      • Bobbo T.
      • Cecchinato A.
      • Bittante G.
      Breed of cow and herd productivity affect milk composition and modeling of coagulation, curd firming, and syneresis.
      ;
      • Schiavon S.
      • Sturaro E.
      • Tagliapietra F.
      • Ramanzin M.
      • Bittante G.
      Nitrogen and phosphorus excretion on mountain farms of different dairy systems.
      ). Briefly, the study area was a mountainous region with a large variability in the characteristics of the farms. To increase the representativeness of the survey, the mixed-breed farms were selected from the different dairy systems spanning from the very traditional extensive farms (e.g., tied animals, summer grazing, winter feeding with hay and concentrates) to the modern intensive ones (e.g., freestalls, TMR, often with silages).
      For this study, generally we sampled 1 farm per week (for 10 consecutive months), taking 1 milk sample per animal from a total of 1,504 cows. The cows comprised purebred Holstein-Friesians (467 cows in 31 herds), Brown Swiss (655 cows in 35 herds), Jerseys (45 cows in 8 herds), Simmentals (158 cows in 20 herds), Rendena (104 cows in 9 herds), and Alpine Greys (75 cows in 14 herds), kept in 41 multibreed herds, corresponding to an average of 37 ± 19 cows sampled per herd-date. The number of herds for each of 15 combinations of breeds was included in a previous paper from the same project (
      • Stocco G.
      • Cipolat-Gotet C.
      • Bobbo T.
      • Cecchinato A.
      • Bittante G.
      Breed of cow and herd productivity affect milk composition and modeling of coagulation, curd firming, and syneresis.
      ). Given that only a limited number of fresh samples could be processed daily for model cheese production in the Milk Laboratory of the Department of Agronomy, Food, Natural Resources, Animals and the Environment (DAFNAE) of the University of Padova (Legnaro, Padova, Italy), for large herds (>50 lactating cows) only a group of preselected cows were sampled, giving preference to breeds that were less well represented and balancing the cows for parities and lactation stages. Milk yield and quality were not taken into account. No cows presenting clinical evidence of disease were sampled.
      Milk samples (2.0 L) were collected from each animal during the evening milking and transported to the DAFNAE laboratory for chemical analysis and individual model cheese-making, as described and discussed by
      • Stocco G.
      • Cipolat-Gotet C.
      • Bobbo T.
      • Cecchinato A.
      • Bittante G.
      Breed of cow and herd productivity affect milk composition and modeling of coagulation, curd firming, and syneresis.
      ,
      • Stocco G.
      • Cipolat-Gotet C.
      • Gasparotto V.
      • Cecchinato A.
      • Bittante G.
      Breed of cow and herd productivity affect milk nutrient recovery in curd, and cheese yield, efficiency and daily production.
      ). From this sample, an aliquot of 1.5 mL was immediately stored at −80°C until the reversed-phase HPLC analysis was performed.

      Milk Composition, Protein Fraction Contents, and Genotype Identification

      A 50-mL aliquot of fresh milk samples was analyzed with a MilkoScan FT6000 (Foss, Hillerød, Denmark) for 3 N compounds (protein, casein, and urea) and for fat and lactose content. Somatic cell counts were obtained from the same aliquot with a Fossomatic FC counter (Foss, Hillerød, Denmark) and converted to SCS according to the formula proposed by
      • Ali A.K.A.
      • Shook G.E.
      An optimum transformation for somatic cell concentration in milk.
      : SCS = 3 + log2 (SCC/100,000).
      The reversed-phase HPLC method described in
      • Maurmayr A.
      • Cecchinato A.
      • Grigoletto L.
      • Bittante G.
      Detection and quantification of α S1-, α S2-, β-, κ-casein, α-lactalbumin, β-lactoglobulin and lactoferrin in bovine milk by reverse-phase high-performance liquid chromatography.
      was used to identify and quantify the contents of αS1-CN, αS2-CN, β-CN, κ-CN, β-LG, and α-LA and to identify the CSN2, CSN3, and BLG genotypes. In the case of κ-CN, it was also possible to distinguish the glycosylated form from the carbohydrate-free form, and total κ-CN was calculated as the sum of these 2 fractions. Briefly, the HPLC apparatus consisted of an Agilent 1260 Infinity Series chromatograph (Agilent Technologies, Santa Clara, CA) equipped with a C8 RP analytical column (Aeris Widepore XB-C8, Phenomenex, Torrance, CA) and a large-pore core-shell packing (3.6 μm, 200 Å, 250 × 2.1 i.d.). A Security Guard Ultra Cartridge System (product no. AJ0-8785, Phenomenex) was used as the pre-column (UHPLC Widepore C8, 2.1 mm i.d.). The following genetic variants were identified: A and B variants of κ-CN; A1, A2, A3, and B variants of β-CN; and A and B variants of β-LG. A detailed description of the method validation can be found in
      • Maurmayr A.
      • Cecchinato A.
      • Grigoletto L.
      • Bittante G.
      Detection and quantification of α S1-, α S2-, β-, κ-casein, α-lactalbumin, β-lactoglobulin and lactoferrin in bovine milk by reverse-phase high-performance liquid chromatography.
      . The frequencies of the alleles of the 3 genes obtained with this method according to breed of cow are reported in Supplemental Figure S1 (https://doi.org/10.3168/jds.2020-18497).
      To better match the 2 sets of data obtained by infrared spectroscopy (CP, casein, and urea expressed in g/L of milk) and by liquid chromatography (individual protein fractions), the HPLC protein fractions were expressed as proportions of the sum of all caseins and multiplied by the milk casein contents obtained using Fourier-transform infrared spectroscopy (FTIR). The detailed nitrogenous profiles of milk in this study consisted of 15 traits; 3 were obtained using FTIR (CP, casein, and urea), 7 were obtained using HPLC (individual protein fractions), and the remaining 5 were calculated either as the sum of 2 of the analyzed traits or the difference between 2 of them, as shown in Supplemental Figure S2 (https://doi.org/10.3168/jds.2020-18497). All the milk nitrogenous fractions were also expressed in percentage of N content of milk, assuming the N content of milk proteins to be 15.67% and that of urea to be 46.67% (Supplemental Figure S2).

      Statistical Analysis

      The data set underwent preliminary editing to exclude samples with residuals outside the range of the mean ± 3 residual standard deviation units for a given trait. The residual standard deviation was obtained using the model described below. The edited data set was analyzed with the MIXED procedure of SAS (SAS Institute Inc., Cary, NC) and the following models:
      yhijkl = μ + herd-dateh + breedi + DIMj + parityk + ehijkl,


      yhijklmno = μ + herd-dateh + breedi + DIMj + parityk + CSN2l + CSN3m + BLGn + ehijklmno,


      where yhijkl and yhijklmno are the protein fractions expressed as grams per liter of milk (g/L) or as percentages of the total N content (%N); μ is the overall mean; herd-dateh is the random effect of the hth herd-date (n = 41); breedi is the fixed effect of the ith breed (6 classes: Holstein-Friesian, Brown Swiss, Jersey, Simmental, Rendena, Alpine Grey); DIMj is the fixed effect of the jth 30-d class of lactation (11 classes); parityk is the fixed effect of the kth class of parity order (k = 1 to ≥4); CSN2l is the fixed effect of the lth class of CSN2 genotype (7 classes: A1A1, A1A2, A2A2, A2A3, BA1, BA2, BB); CSN3m is the fixed effect of the mth class of CSN3 genotype (3 classes: AA, AB, BB); BLGn is the fixed effect of the nth class of BLG genotype (3 classes: AA, AB, BB); and ehijkl and ehijklmno are the residual random errors ∼N (0, σe2). As first step, both the models were treated as random effects models (all the effects were considered to be random) for estimating the variance components and calculating their proportions on total variance. As a second step, the 2 models were treated as mixed models for estimating the least squares means of the effect of breed. In this latter case, the effects of genotype were included in the models with the sole aim of comparing the estimates of the effects of breed with and without this genetic information.
      The following orthogonal contrasts (P < 0.05) were used to compare the different breeds: (1) specialized breeds (Holstein-Friesian + Brown Swiss + Jersey) versus dual-purpose breeds (Simmental + Rendena + Alpine Grey); (2) within specialized breeds: the 2 large-framed breeds (Holstein-Friesian and Brown Swiss) versus the small-framed breed (Jersey); (3) within the 2 large-framed specialized breeds: Holstein-Friesian versus Brown Swiss; (4) within the dual-purpose breeds: the large-framed cosmopolitan breed (Simmental) versus the 2 medium-sized local breeds (Rendena + Alpine Grey); and (5) within the 2 medium-sized local breeds: Rendena versus Alpine Grey.

      RESULTS AND DISCUSSION

      Detailed Protein Profiles of Bovine Milk

      Descriptive statistics of the detailed milk protein quantitative (g/L milk) and qualitative (%N) profiles are summarized in Table 1, together with daily milk yield, fat and lactose contents, and SCS. Milk yield and composition were reported and discussed in a previous study on milk coagulation and curd firming properties (
      • Stocco G.
      • Cipolat-Gotet C.
      • Bobbo T.
      • Cecchinato A.
      • Bittante G.
      Breed of cow and herd productivity affect milk composition and modeling of coagulation, curd firming, and syneresis.
      ). On average, the milk had a relatively large CP content (36 g/L) with a coefficient of variation (CV) of 13 g/L. About 93% of CP (analyzed by FTIR) consisted of the true proteins (analyzed by HPLC), with the remaining 7% of N identified as NPN compounds. The standard deviation of their proportions (2.95% of total N for both traits, one being the complement to 100% of the other) represented a CV of 3% when calculated on the average value of the true proteins and 40% when calculated on the average value of the NPN compounds (Table 1). The CV of the true proteins and NPN compounds expressed as grams per liter increased to 13 and 44%, respectively, because of the additional variability arising from multiplying the %N values by the protein content of the milk (in g/L). Similarly, the CV of the quantitative data (expressed in g/L) is greater than the CV of the qualitative data (in %N) for all the other nitrogenous fractions of milk (Table 1). These values are close to the results shown by
      • Gustavsson F.
      • Buitenhuis A.J.
      • Johansson M.
      • Bertelsen H.P.
      • Glantz M.
      • Poulsen N.A.
      • Lindmark Månsson H.
      • Stålhammar H.
      • Larsen L.B.
      • Bendixen C.
      • Paulsson M.
      • Andrén A.
      Effects of breed and casein genetic variants on protein profile in milk from Swedish red, Danish Holstein, and Danish Jersey cows.
      on 3 Scandinavian breeds, where the sum of the single protein fractions represented 94% of the total CP. The discussion of the results of detailed protein profiles is based only on studies using gold standard methods (HPLC and capillary zone electrophoresis) because the predictions based on secondary methods, mainly infrared spectroscopic techniques, are characterized by moderate accuracy and their variability often results from the variability of milk spectra more than that of the reference values. Moreover, in the case of multibreed studies, genotypes of milk fractions are generally not known and not included in the statistical models.
      Table 1Descriptive statistics for milk yield, milk composition, and protein composition (n = 1,504 cows)
      TraitMeanSDPercentile
      5th95th
      Milk yield, kg/d24.289.0611.4039.90
      Fat, %4.010.952.855.42
      Lactose, %4.850.244.415.15
      SCS, units3.081.940.316.64
      Protein composition, g/L of milk
       CP36.154.6429.0044.50
       True protein33.754.2727.1741.43
      Caseins
      Sum of caseins: αS1-CN + αS2-CN + β-CN+ κ-CN.
      28.363.5323.0034.80
      αS1-CN9.411.367.3711.77
      αS2-CN2.780.611.743.82
      β-CN10.691.328.6813.02
      κ-CN
      Glycosylated κ-CN + carbohydrate-free κ-CN.
      5.451.223.457.52
      Glycosylated κ-CN2.040.780.933.51
      Carbohydrate-free κ-CN3.380.732.234.61
      Whey proteins
      Sum of whey proteins: β-LG + α-LA.
      5.421.203.487.48
      β-LG4.591.172.666.60
      α-LA0.760.130.550.97
       NPN compounds
      Sum of urea + minor NPN compounds (e.g., small peptides, ammonia, creatine, creatinine).
      2.491.090.894.46
      MUN0.250.090.120.43
      Minor NPN compounds1.751.020.253.56
      Protein composition, % of total milk N
       True protein93.462.9588.4298.20
      Caseins78.491.1476.6080.29
      αS1-CN26.161.8023.4729.14
      αS2-CN7.701.375.139.74
      β-CN29.692.2725.8533.33
      κ-CN15.022.4010.7618.66
       Glycosylated κ-CN5.631.862.778.90
       Carbohydrate-free κ-CN9.341.646.4111.88
       Whey proteins14.982.8010.3119.47
       β-LG12.672.767.9017.09
       α-LA2.110.391.482.80
       NPN compounds6.832.752.6611.62
       MUN2.050.770.943.55
       Minor NPN compounds4.762.620.729.36
      1 Sum of caseins: αS1-CN + αS2-CN + β-CN+ κ-CN.
      2 Glycosylated κ-CN + carbohydrate-free κ-CN.
      3 Sum of whey proteins: β-LG + α-LA.
      4 Sum of urea + minor NPN compounds (e.g., small peptides, ammonia, creatine, creatinine).
      The NPN compounds consisted of MUN, about 2% of all milk N, corresponding to 0.25 g/L urea, and other minor NPN compounds, around 4.8% of all milk N. The levels of MUN were consistent with those often found in different farming systems and breeds (
      • Bastin C.
      • Laloux L.
      • Gillon A.
      • Miglior F.
      • Soyeurt H.
      • Hammami H.
      • Bertozzi C.
      • Gengler N.
      Modeling milk urea of Walloon dairy cows in management perspectives.
      ). The true proteins consisted of the major casein fractions (28.36 g/L) and whey proteins (5.42 g/L), corresponding to about 78 and 15% of milk CP, respectively. In the work of
      • Gustavsson F.
      • Buitenhuis A.J.
      • Johansson M.
      • Bertelsen H.P.
      • Glantz M.
      • Poulsen N.A.
      • Lindmark Månsson H.
      • Stålhammar H.
      • Larsen L.B.
      • Bendixen C.
      • Paulsson M.
      • Andrén A.
      Effects of breed and casein genetic variants on protein profile in milk from Swedish red, Danish Holstein, and Danish Jersey cows.
      , they represented 83 and 11%, respectively. As expected, β-CN and αS1-CN were the main casein fractions, with average contents of 10.7 g/L (30% N) and 9.4 g/L (26% N), respectively. The contents of αS2-CN (2.8 g/L; 8% N) and κ-CN (5.5 g/L; 15% N; represented more by its carbohydrate-free form than by its glycosylated form) were much lower than the other 2 casein fractions. In the case of the whey proteins, the β-LG content (4.6 g/L; 13% N) was about 6 times greater than the α-LA content (0.8 g/L; 2% N), with CV of 25 and 17%, respectively. In general, the proportions within caseins and whey proteins confirmed those found by
      • Gustavsson F.
      • Buitenhuis A.J.
      • Johansson M.
      • Bertelsen H.P.
      • Glantz M.
      • Poulsen N.A.
      • Lindmark Månsson H.
      • Stålhammar H.
      • Larsen L.B.
      • Bendixen C.
      • Paulsson M.
      • Andrén A.
      Effects of breed and casein genetic variants on protein profile in milk from Swedish red, Danish Holstein, and Danish Jersey cows.
      , with some differences in their concentrations probably due to the different breeds studied and to their contribution to the total population.

      Effects of Herd and Date, Parity, and Lactation Stage on the Detailed Protein Profiles of Milk

      Figure 1 shows the proportions of the total phenotypic variance explained by the major sources of variation included in the statistical models for each protein fraction expressed qualitatively (%N) and quantitatively (g/L milk). As can be seen, almost every protein fraction has its particular pattern of variation. Given that very few studies, as far as we know, have reported this kind of information, and none have compared different breeds, direct comparisons with previous research are not possible.
      • Bobe G.
      • Beitz D.C.
      • Freeman A.E.
      • Lindberg G.L.
      Effect of milk protein genotypes on milk protein composition and its genetic parameter estimates.
      and
      • Schopen G.C.B.
      • Heck J.M.L.
      • Bovenhuis H.
      • Visker M.H.P.W.
      • van Valenberg H.J.F.
      • van Arendonk J.A.M.
      Genetic parameters for major milk proteins in Dutch Holstein-Friesians.
      analyzed several sources of variability in the major milk protein fractions, including the effects of the genotypes of the CSN2, CSN3, and BLG genes, but only in the Holstein breed.
      Figure thumbnail gr1
      Figure 1Proportion of the phenotypic variance of the protein fractions expressed as proportion of total milk nitrogen (%N) and as content of milk (g/L) explained by the effects of the models before (base model) and after (+Gen model) the inclusion of the genotypes of β-CN, κ-CN, and β-LG. TP = true protein, Glyco-κ-CN = glycosylated κ-CN; CF-κ-CN = carbohydrate-free κ-CN; WP = whey protein.
      Investigation of the effects of herd-date and of the cows' parity and lactation stage is not among the aims of this research and is therefore not reported or discussed in detail. They were included in the models to reduce possible confounding effects and to obtain an unbiased estimation of the other effects. Briefly, the herd-date effect (a factor common to all cows sampled on the same farm and the same date) ranged, with both models, from about 5% of total variance for αS2-CN, κ-CN, and minor NPN compounds to about 20% for the sum of casein fractions and α-LA and peaked at the very high proportion of all variance, about 75%, for MUN (
      • Dadousis C.
      • Cipolat-Gotet C.
      • Schiavon S.
      • Bittante G.
      • Cecchinato A.
      Inferring individual cow effects, dairy system effects and feeding effects on latent variables underlying milk protein composition and cheese-making traits in dairy cattle.
      ). This clearly confirms the significance of urea as an environmental indicator, given that the content of urea in milk, such as urea in blood and ammonia in the rumen fluid, depends mainly on the feeding strategies adopted by the different farms and especially on the level of CP in the diet (
      • Aguilar M.
      • Hanigan M.D.
      • Tucker H.A.
      • Jones B.L.
      • Garbade S.K.
      • McGilliard M.L.
      • Stallings C.C.
      • Knowlton K.F.
      • James R.E.
      Cow and herd variation in milk urea nitrogen concentrations in lactating dairy cattle.
      ;
      • Schiavon S.
      • Cesaro G.
      • Tagliapietra F.
      • Gallo L.
      • Bittante G.
      Influence of N shortage and conjugated linoleic acid supplementation on some productive, digestive, and metabolic parameters of lactating cows.
      ). Herd-date represented almost 14% of total variance in the case of the CP content of milk, a value similar to other results found in the scientific literature and lower than the variation in daily milk yield (
      • Stocco G.
      • Cipolat-Gotet C.
      • Bobbo T.
      • Cecchinato A.
      • Bittante G.
      Breed of cow and herd productivity affect milk composition and modeling of coagulation, curd firming, and syneresis.
      ). Herd-date had little influence on the minor milk NPN compounds, and among the other protein fractions it had the greatest effect on α-LA. Similar results were reported by
      • Schopen G.C.B.
      • Heck J.M.L.
      • Bovenhuis H.
      • Visker M.H.P.W.
      • van Valenberg H.J.F.
      • van Arendonk J.A.M.
      Genetic parameters for major milk proteins in Dutch Holstein-Friesians.
      , who found that herd explained a similar or higher proportion of phenotypic variance in protein fractions, although they also found that α-LA was the protein fraction most affected by herd. It is worth noting that this same survey showed that herd-date appeared to have a similar effect on the content of serum albumins in the blood of the same cows (greater than the effect on the serum globulin content) and that it was greater in the herds with the highest daily milk yield (
      • Bobbo T.
      • Fiore E.
      • Gianesella M.
      • Morgante M.
      • Gallo L.
      • Ruegg P.L.
      • Bittante G.
      • Cecchinato A.
      Variation in blood serum proteins and association with somatic cell count in dairy cattle from multi-breed herds.
      ), confirming results obtained by
      • Rowlands G.J.
      • Little W.
      • Kitchenham B.A.
      Relationships between blood composition and fertility in dairy cows—A field study.
      . An increase in dietary CP also resulted in increased serum protein (
      • Hoffman P.C.
      • Esser N.M.
      • Bauman L.M.
      • Denzine S.L.
      • Engstrom M.
      • Chester-Jones H.
      Short communication: Effect of dietary protein on growth and nitrogen balance of Holstein heifers.
      ;
      • Raggio G.
      • Lobley G.E.
      • Berthiaume R.
      • Pellerin D.
      • Allard G.
      • Dubreuil P.
      • Lapierre H.
      Effect of protein supply on hepatic synthesis of plasma and constitutive proteins in lactating dairy cows.
      ;
      • Law R.A.
      • Young F.J.
      • Patterson D.C.
      • Kilpatrick D.J.
      • Wylie A.R.G.
      • Mayne C.S.
      Effect of dietary protein content on animal production and blood metabolites of dairy cows during lactation.
      ).
      Although the effects of parity and DIM were significant for almost all the qualitative and quantitative milk protein profile traits (Table 2), the proportion of total phenotypic variance explained by parity was always low (≤7% of total phenotypic variance) for all the qualitative (%N) and quantitative (g/L) traits (
      • Ng-Kwai-Hang K.F.
      • Hayes J.F.
      • Moxley J.E.
      • Monardes H.G.
      Variation in milk protein concentrations associated with genetic polymorphism and environmental factors.
      ), whereas the proportion explained by lactation stage was low in the case of the qualitative traits (%N), except carbohydrate-free κ-CN and α-LA, and much greater for the majority of the quantitative protein traits (g/L). In fact, the milk content of CP and all the major protein fractions tended, as expected, to increase in mid and late lactation (data not shown). Only the quantitative fractions present in milk in small amounts (αS2-CN, glycosylated κ-CN, α-LA, NPN compounds, MUN, and minor NPN compounds) were not much affected by stage of lactation.
      Table 2Results from linear mixed models based on herd-date, breed, parity, and DIM (base model) or with the inclusion of the genotypes (+Gen) of CSN2, CSN3, and BLG genes (F-value and significance) for qualitative (% of total milk N; %N) and quantitative (g/L of milk) detailed milk protein profile
      TraitUnitModelFixed effect (F-value)RMSE
      Root mean squared error, expressed as percentage of the total milk.
      BreedDIMParityCSN2CSN3BLG
      CPg/LBase47.1
      P < 0.001.
      75.1
      P < 0.001.
      10.1
      P < 0.001.
      3.00
      +Gen32.4
      P < 0.001.
      71.2
      P < 0.001.
      10.3
      P < 0.001.
      5.0
      P < 0.001.
      0.61.23.00
      True protein%NBase11.7
      P < 0.001.
      0.90.32.64
      +Gen3.0
      P < 0.01
      1.61.025.1
      P < 0.001.
      14.2
      P < 0.001.
      381.8
      P < 0.001.
      1.94
      g/LBase38.5
      P < 0.001.
      74.8
      P < 0.001.
      10.2
      P < 0.001.
      2.80
      +Gen31.4
      P < 0.001.
      75.9
      P < 0.001.
      12.1
      P < 0.001.
      4.5
      P < 0.001.
      0.117.3
      P < 0.001.
      2.72
      Caseins
      Sum of caseins: αS1-CN + αS2-CN + β-CN+ κ-CN.
      %NBase2.8
      P < 0.05
      11.3
      P < 0.001.
      35.8
      P < 0.001.
      0.92
      +Gen3.0
      P < 0.05
      10.3
      P < 0.001.
      30.7
      P < 0.001.
      0.24.8
      P < 0.01
      3.2
      P < 0.05
      0.92
      g/LBase47.4
      P < 0.001.
      70.0
      P < 0.001.
      14.5
      P < 0.001.
      2.33
      +Gen32.8
      P < 0.001.
      66.4
      P < 0.001.
      14.8
      P < 0.001.
      5.8
      P < 0.001.
      0.90.82.30
      αS1-CN%NBase30.1
      P < 0.001.
      3.0
      P < 0.01
      0.31.56
      +Gen7.3
      P < 0.001.
      4.3
      P < 0.001.
      1.73.8
      P < 0.001.
      175.6
      P < 0.001.
      0.61.33
      g/LBase13.4
      P < 0.001.
      37.8
      P < 0.001.
      8.9
      P < 0.001.
      1.00
      +Gen32.1
      P < 0.001.
      39.3
      P < 0.001.
      11.7
      P < 0.001.
      4.3
      P < 0.001.
      49.6
      P < 0.001.
      1.10.93
      αS2-CN%NBase14.0
      P < 0.001.
      6.1
      P < 0.001.
      10.4
      P < 0.001.
      1.24
      +Gen7.2
      P < 0.001.
      6.8
      P < 0.001.
      10.7
      P < 0.001.
      15.0
      P < 0.001.
      24.8
      P < 0.001.
      0.21.11
      g/LBase35.8
      P < 0.001.
      5.3
      P < 0.001.
      14.8
      P < 0.001.
      0.50
      +Gen15.6
      P < 0.001.
      7.0
      P < 0.001.
      15.8
      P < 0.001.
      8.0
      P < 0.001.
      19.9
      P < 0.001.
      0.10.47
      β-CN%NBase46.4
      P < 0.001.
      6.7
      P < 0.001.
      11.6
      P < 0.001.
      1.76
      +Gen5.7
      P < 0.001.
      8.0
      P < 0.001.
      17.8
      P < 0.001.
      6.6
      P < 0.001.
      137.4
      P < 0.001.
      3.01.49
      g/LBase3.0
      P < 0.05
      41.7
      P < 0.001.
      15.3
      P < 0.001.
      1.05
      +Gen17.6
      P < 0.001.
      46.0
      P < 0.001.
      19.5
      P < 0.001.
      7.1
      P < 0.001.
      35.3
      P < 0.001.
      1.11.00
      κ-CN
      Glycosylated κ-CN + carbohydrate-free κ-CN.
      %NBase54.0
      P < 0.001.
      4.3
      P < 0.001.
      8.2
      P < 0.001.
      2.03
      +Gen6.9
      P < 0.001.
      8.8
      P < 0.001.
      23.9
      P < 0.001.
      2.4
      P < 0.05
      602.0
      P < 0.001.
      0.11.37
      g/LBase94.9
      P < 0.001.
      38.7
      P < 0.001.
      2.30.87
      +Gen13.9
      P < 0.001.
      58.5
      P < 0.001.
      5.3
      P < 0.01
      4.4
      P < 0.001.
      326.3
      P < 0.001.
      0.90.67
      Glycosylated κ-CN%NBase11.7
      P < 0.001.
      21.8
      P < 0.001.
      13.3
      P < 0.001.
      1.61
      +Gen9.4
      P < 0.001.
      32.6
      P < 0.001.
      25.6
      P < 0.001.
      1.5239.4
      P < 0.001.
      0.21.32
      g/LBase27.0
      P < 0.001.
      44.4
      P < 0.001.
      10.5
      P < 0.001.
      0.62
      +Gen5.3
      P < 0.001.
      61.1
      P < 0.001.
      18.3
      P < 0.001.
      3.1
      P < 0.01
      190.0
      P < 0.001.
      0.20.53
      Carbohydrate-free κ-CN%NBase60.9
      P < 0.001.
      11.4
      P < 0.001.
      5.4
      P < 0.01
      1.35
      +Gen14.8
      P < 0.001.
      13.6
      P < 0.001.
      5.2
      P < 0.01
      0.9122.5
      P < 0.001.
      0.71.19
      g/LBase103.1
      P < 0.001.
      5.5
      P < 0.001.
      3.4
      P < 0.05
      0.56
      +Gen29.7
      P < 0.001.
      5.3
      P < 0.001.
      3.1
      P < 0.05
      1.793.3
      P < 0.001.
      1.70.50
      Whey proteins
      Sum of whey proteins: β-LG + α-LA.
      %NBase15.6
      P < 0.001.
      0.87.7
      P < 0.001.
      2.47
      +Gen7.9
      P < 0.001.
      1.814.3
      P < 0.001.
      32.8
      P < 0.001.
      23.8
      P < 0.001.
      428.9
      P < 0.001.
      1.74
      g/LBase12.2
      P < 0.001.
      21.8
      P < 0.001.
      4.4
      P < 0.01
      0.96
      +Gen13.1
      P < 0.001.
      36.9
      P < 0.001.
      7.6
      P < 0.001.
      18.9
      P < 0.001.
      10.5
      P < 0.001.
      270.9
      P < 0.001.
      0.76
      β-LG%NBase16.4
      P < 0.001.
      1.9
      P < 0.05
      7.8
      P < 0.001.
      2.44
      +Gen7.5
      P < 0.001.
      4.9
      P < 0.001.
      14.1
      P < 0.001.
      34.8
      P < 0.001.
      14.7
      P < 0.001.
      448.9
      P < 0.001.
      1.70
      g/LBase13.8
      P < 0.001.
      22.2
      P < 0.001.
      6.2
      P < 0.001.
      0.94
      +Gen14.0
      P < 0.001.
      40.4
      P < 0.001.
      9.7
      P < 0.001.
      23.6
      P < 0.001.
      8.4
      P < 0.001.
      307.7
      P < 0.001.
      0.72
      α-LA%NBase19.8
      P < 0.001.
      23.8
      P < 0.001.
      1.10.30
      +Gen11.4
      P < 0.001.
      30.8
      P < 0.001.
      0.845.3
      P < 0.001.
      40.4
      P < 0.001.
      0.60.26
      g/LBase8.8
      P < 0.001.
      0.97.6
      P < 0.001.
      0.11
      +Gen8.8
      P < 0.001.
      0.87.8
      P < 0.001.
      26.7
      P < 0.001.
      30.6
      P < 0.001.
      0.80.10
      NPN compounds
      Sum of urea + minor NPN compounds (e.g., small peptides, ammonia, creatine, creatinine).
      %NBase11.6
      P < 0.001.
      0.90.72.45
      +Gen3.1
      P < 0.01
      1.41.822.5
      P < 0.001.
      12.6
      P < 0.001.
      372.8
      P < 0.001.
      1.80
      g/LBase16.4
      P < 0.001.
      4.6
      P < 0.001.
      0.80.94
      +Gen8.6
      P < 0.001.
      6.6
      P < 0.001.
      1.320.3
      P < 0.001.
      10.4
      P < 0.001.
      315.8
      P < 0.001.
      0.70
      MUN%NBase3.9
      P < 0.01
      10.7
      P < 0.001.
      3.3
      P < 0.05
      0.38
      +Gen3.9
      P < 0.01
      9.8
      P < 0.001.
      2.6
      P < 0.05
      1.90.60.20.38
      g/LBase7.0
      P < 0.001.
      4.2
      P < 0.001.
      5.8
      P < 0.001.
      0.04
      +Gen2.6
      P < 0.05
      3.8
      P < 0.001.
      4.9
      P < 0.01
      0.41.20.20.04
      Minor NPN compounds%NBase12.6
      P < 0.001.
      1.50.32.43
      +Gen3.7
      P < 0.01
      2.7
      P < 0.01
      1.125.1
      P < 0.001.
      13.3
      P < 0.001.
      385.5
      P < 0.001.
      1.76
      g/LBase13.8
      P < 0.001.
      4.7
      P < 0.001.
      0.70.94
      +Gen5.7
      P < 0.001.
      7.4
      P < 0.001.
      1.120.9
      P < 0.001.
      10.8
      P < 0.001.
      321.0
      P < 0.001.
      0.70
      1 Root mean squared error, expressed as percentage of the total milk.
      2 Sum of caseins: αS1-CN + αS2-CN + β-CN+ κ-CN.
      3 Glycosylated κ-CN + carbohydrate-free κ-CN.
      4 Sum of whey proteins: β-LG + α-LA.
      5 Sum of urea + minor NPN compounds (e.g., small peptides, ammonia, creatine, creatinine).
      * P < 0.05
      ** P < 0.01
      *** P < 0.001.
      Inclusion of the genotypes of CSN2, CSN3, and BLG in the model used to analyze the qualitative and quantitative detailed milk protein profiles made little difference to the relative importance of herd-date, parity, and DIM (see Figure 1). This is, of course, expected because the frequency of the genotypes within breed does not differ much across different farms and because a cow's genotypes do not change during its life.

      Proportion of the Phenotypic Variance in the Detailed Milk Protein Profiles Explained by Breed of Cow

      The effect of breed of cow, which is the most important permanent individual factor, on the proportions of the detailed protein fractions in total milk N (%N) was always significant when treated as a fixed factor (Table 2). When breed was considered a random factor, we were able to estimate that, with the exception of the sum of caseins and of MUN, it had a similar or greater effect on total variance than herd-date (Figure 1, base model, %N). In the case of urea, the proportion of which %N is greatly affected by herd-date, breed within the same herd-date was of almost no importance, confirming again that it depends more on the farm's dietary regimen and on rumen metabolism than on the cow's genetics.
      The finding that individual caseins, but not their sum, were highly affected by breed of cow suggests that the ratio of the caseins to the other nitrogenous compounds, including whey proteins, does not vary much. In fact, while keeping the same casein ratio, some individual caseins can be partially replaced by others according to the genetics of the cow, especially its breed (Figure 1, base model, %N).
      Breed greatly affected the total variability in the CP content of milk (23% of total phenotypic variance), which explains why the effect of breed on quantitatively measured protein fractions (Figure 1, base model, g/L) is often larger than when they are qualitatively measured (Figure 1, base model, %N). The most notable exception was β-CN: breed was an important source of variation (almost 23%) when the casein fraction was expressed as %N but negligible (<1%) when expressed as grams per liter of milk. We confirmed that breed did not much affect the variability in the milk content of urea or its proportion of total milk N. However, the effect of breed was significant for all the quantitative traits (Table 2).
      The final result regarding the patterns in the major sources of variation (Figure 1, base model, %N) was that the residual variance represented the major source of variation in all the protein traits expressed in %N, ranging from 50 to >80% of total variance, with the notable exception of milk urea (22%). The residual variance was also the most important source of variation in the entire quantitative milk protein profile, except for urea, but it was less important than in the corresponding qualitative traits (Figure 1, base model, g/L vs. Figure 1, base model, %N). It is worth bearing in mind that the residual variance includes the effects of all the common, individual, permanent, and temporary effects not included in the statistical model and, notably, all the genetic effects not represented by the cow's breed.

      Proportion of the Phenotypic Variance in the Detailed Milk Protein Profiles Explained by the CSN2, CSN3, and BLG Genotypes

      It is well known that the genes encoding for the casein fractions are in tight genetic linkage within a 250-kb region on chromosome 6. This results in the cotransmission of casein gene variants across generations and in a low recombination rate between them. Casein genotypes are therefore not independent, and it is not easy to estimate the effect of the single gene (
      • Martin P.
      • Szymanowska M.
      • Zwierzchowski L.
      • Leroux C.
      The impact of genetic polymorphisms on the protein composition of ruminant milks.
      ;
      • Caroli A.M.
      • Chessa S.
      • Erhardt G.J.
      Invited review: Milk protein polymorphisms in cattle: Effect on animal breeding and human nutrition.
      ). However, the aim of this work was not to estimate or evaluate the effect of the major genes alleles. Instead, the aim was to address the importance of these genes in explaining the phenotypic variability of animals and the differences between the breeds.
      Figure 1 (+Gen model, %N) shows the proportions of the total phenotypic variance explained by the major sources of variation together with the genotypes of the CSN2, CSN3, and BLG genes included in the statistical models of each protein fraction expressed as qualitative traits (%N). As can be seen, inclusion of the genotypes of the 3 milk protein loci dramatically changed the relative importance of the different sources of variation in all the qualitative traits except total caseins and milk urea. It is worth noting here that numerous studies have examined the effect of individual genetic variants on many milk traits, but the authors are unaware of studies quantifying the effects of genotypes on sources of variation of milk protein profile, so no direct comparison with previous research can be made.
      Comparison between Figure 1, base model, %N and Figure 1, +Gen model, %N shows that the CSN2 gene, which encodes for β-CN, exerted the smallest effects compared with the other 2 loci. The genotypes of this gene (A1A1, A1A2, A2A2, A2A3, BA1, BA2, and BB), despite significantly affecting the majority of the qualitative milk protein traits (Table 2), exerted an appreciable effect (>10% of total variance) on only 1 casein fraction: curiously, αS2-CN and not β-CN. In contrast, they affected all the noncasein traits, except urea, and the milk true protein (Figure 1, +Gen model, %N). In the study by
      • Schopen G.C.B.
      • Heck J.M.L.
      • Bovenhuis H.
      • Visker M.H.P.W.
      • van Valenberg H.J.F.
      • van Arendonk J.A.M.
      Genetic parameters for major milk proteins in Dutch Holstein-Friesians.
      , the β-CN genotype explained a lower proportion of phenotypic variance than in ours, but it still explained a higher proportion of variance in αS2-CN than β-CN did. Note that, in the case of the qualitative data, NPN compounds and true protein together come to 100%, so the proportion of one trait can be obtained by subtracting the proportions of the other traits from 100%. This means that factors affecting one trait exert exactly the same effect, with opposite sign, on the other trait. Moreover, minor NPN compounds account for the greater part of NPN compounds, so it is not unexpected that they follow a similar pattern.
      The genotypes of the BLG gene (AA, AB, and BB), which encode for β-LG, had a much greater effect (>40% of total variance) on all noncasein traits except α-LA and MUN.
      • Schopen G.C.B.
      • Heck J.M.L.
      • Bovenhuis H.
      • Visker M.H.P.W.
      • van Valenberg H.J.F.
      • van Arendonk J.A.M.
      Genetic parameters for major milk proteins in Dutch Holstein-Friesians.
      found that the genotypes of the BLG gene explained 70% of the total phenotypic variance in β-LG expressed as percentage of total protein.
      The reverse pattern was observed for the genotypes of CSN3 (AA, AB, and BB), which encode for κ-CN: they significantly affected almost all the qualitative traits, but the effect was notable only for the individual casein fractions, except αS2-CN (Figure 1, +Gen model, %N). The effect was particularly marked in the case of the proportion of κ-CN out of the total milk N content, which explained 70% of total variance. Other authors (
      • Bobe G.
      • Beitz D.C.
      • Freeman A.E.
      • Lindberg G.L.
      Effect of milk protein genotypes on milk protein composition and its genetic parameter estimates.
      ;
      • Schopen G.C.B.
      • Heck J.M.L.
      • Bovenhuis H.
      • Visker M.H.P.W.
      • van Valenberg H.J.F.
      • van Arendonk J.A.M.
      Genetic parameters for major milk proteins in Dutch Holstein-Friesians.
      ) also found CSN3 to have a notable effect on κ-CN content.
      The large proportion of phenotypic variance explained by inclusion of the CSN2, CSN3, and BLG genotypes in the models was accompanied by a parallel drastic reduction in the proportions explained by breed and the residual, except for total casein and urea (compare Figures 1, +Gen model, %N and Figure 1, base model, %N). This means that the effects of these 3 major genes represent (1) a notable portion of the total phenotypic variance in the detailed qualitative milk protein profile, (2) the greatest part of the differences between the detailed protein profiles of milk from different breeds, and (3) an important part of the genetic variation among different cows within the same breed.
      • Hallén E.
      • Wedholm A.
      • Andrén A.
      • Lundén A.
      Effect of β-casein, κ-casein and β-lactoglobulin genotypes on concentration of milk protein variants.
      also found that including the haplotypes of the CSN2/CSN3 gene cluster and of the genotypes of the BLG gene resulted in a large reduction in the residual variance.
      Moving from the qualitative to the quantitative data (in g/L), we noted that the CP content of milk was not much affected by the genotypes of the 3 major genes tested (Figure 1, +Gen model, g/L). The content of true protein in milk depends much more on the CP content (g/L) than on its proportion of total N (%N). This explains why the effect of the genotypes of the 3 major genes on the content of true protein in milk was negligible but notable when true protein was expressed as percentage of total milk N. The casein content of milk, being highly correlated with the crude and true protein contents, also seems to be not much affected by the genotypes of CSN2, CSN3, and BLG (Figure 1, +Gen model, g/L).
      The effects of the 3 major genes on the content of all the nitrogenous fractions (excluding urea) in milk were notable, although slightly less so than in the case of the qualitative data (%N). The pattern was also similar: CSN2 had a greater effect on the contents of αS2-CN and the noncasein traits than on the contents of β-CN and the other caseins; CSN3 influenced the milk content of all the individual casein fractions (especially the κ-CN fractions) as well as α-LA; and BLG had a notable effect on the content of all the noncasein fractions, except α-LA and urea. These results are in line with those of
      • Bobe G.
      • Beitz D.C.
      • Freeman A.E.
      • Lindberg G.L.
      Effect of milk protein genotypes on milk protein composition and its genetic parameter estimates.
      , who found that the CSN3 genotypes, in Holstein cows, mainly influenced κ-CN and αS1-CN, whereas the BLG genotypes mainly affected β-LG itself but also some of the caseins.
      • Heck J.M.L.
      • Schennink A.
      • Van Valenberg H.J.F.
      • Bovenhuis H.
      • Visker M.H.P.W.
      • Van Arendonk J.A.M.
      • Van Hooijdonk A.C.M.
      Effects of milk protein variants on the protein composition of bovine milk.
      found that 90% of the genetic variation in the proportion of β-LG can be explained by its genotypes.
      In the case of the quantitative milk protein profiles, the proportion of total variance explained by breed of cow and by the residuals also changed considerably after including the genotypes, although there were some differences from the qualitative data. First of all, the importance of breed was little altered by inclusion of the genotypes in the model in the case of the milk contents of CP, true protein, casein, and urea (compare Figure 1, +Gen model, g/L and Figure 1, base model, g/L). In the case of the NPN traits, as with the qualitative data, the differences among breeds tended to decrease after correcting for the genotypes of the major genes. For example, part of the differences among breeds depends on the different frequencies of the alleles of the 3 genes examined. In some cases, however, the results were very different. With respect to the milk content of αS1-CN, the proportion of total variance explained by breed more than doubled after including the genotypes in the model. The difference was even more evident with the milk content of β-CN, where the effect of breed appeared to be almost null in the base model but became evident (12%) after inclusion of the genotypes, and when added to the effects of these 3 genes, it represented about a quarter of the total variance. This could be interpreted as both the major genes and the polygene affecting the milk content of β-CN, but differently. The breeds exhibiting a favorable effect of the polygenes are probably those with higher frequencies of the unfavorable alleles of the major genes.

      Detailed Protein Profiles of Milk from Specialized Breeds Versus Dual-Purpose Breeds

      The orthogonal contrast between the 3 specialized dairy breeds (Holstein-Friesian, Brown Swiss, and Jersey) and the 3 dual-purpose breeds (Simmental, Rendena, and Alpine Grey) was significant for 8 out of 14 of the qualitative milk protein traits (in %N) with the base model (herd-date, breed, parity, and DIM) and for 10 traits with the +Gen model (also including the genotypes of CNS2, CNS3, and BLG). In the case of the 15 quantitative milk protein traits (in g/L), the orthogonal contrast was significant for 10 traits with the base model and 9 with the +Gen model. The large number of traits reaching statistical significance is mainly due to the large amount of data analyzed rather than to notable differences between the 2 groups of breeds. As shown in the following paragraphs, by analyzing 6 breeds, we were able to reveal much more notable differences among the breeds in each of the 2 groups than between the 2 groups. These latter contrasts are therefore not described or discussed in detail, also because of the lack of studies comparing the detailed protein profiles obtained by gold standard laboratory methods from the milk of cows of several breeds determined both quantitatively and qualitatively, with or without inclusion of the effects of some major milk protein genes. Therefore, the discussion is limited to studies that deal with some of the traits of a few breeds, often without parallel consideration of the effects of major genes.

      Detailed Protein Profiles of Milk from Jersey Versus Holstein-Friesian + Brown Swiss Cows

      Within the specialized dairy breeds, the first comparison we make is between the small-framed Jersey breed and the average of the 2 large-framed breeds, Holstein-Friesian and Brown Swiss. The results should be taken with prudence because of the relatively low numbers of Jersey cows compared with Holstein-Friesian and Brown Swiss cows. However, our results were consistent with the literature. We confirmed the Jersey breed's well-known characteristic of producing milk with a high CP content (
      • Bland J.H.
      • Grandison A.S.
      • Fagan C.C.
      Effect of blending Jersey and Holstein-Friesian milk on Cheddar cheese processing, composition, and quality.
      ;
      • Stocco G.
      • Cipolat-Gotet C.
      • Bobbo T.
      • Cecchinato A.
      • Bittante G.
      Breed of cow and herd productivity affect milk composition and modeling of coagulation, curd firming, and syneresis.
      )—much higher than that of the other 2 specialized dairy breeds (Figure 2). This was reflected in much greater contents of true protein and casein (Figure 2) and of all the individual casein fractions in the milk (Figure 3, Figure 4). Comparing the Jerseys with the 2 large-framed breeds, statistical significance failed to be reached only in the case of β-CN. In contrast, Jersey milk had significantly smaller β-LG and whey protein contents but nonsignificantly smaller α-LA contents than milk from the larger cows (Figure 5); it also had greater contents of the NPN compounds, except urea N (Figure 6).
      Figure thumbnail gr2
      Figure 2Breed effect of CP, true protein, and total casein expressed as content of milk (g/L) and as proportion of total milk nitrogen (%N) estimated according to a model including herd-date, parity, and lactation stage (base model, in solid color) or also the genotypes of CSN2, CSN3, and BLG (+Gen model, in dashed color). Asterisks (*P < 0.05; **P < 0.01; ***P < 0.001) on the Jersey bar refer to the orthogonal contrast Jersey versus (Holstein + Brown Swiss); on Holstein bar for the orthogonal contrast Holstein versus Brown Swiss; on Simmental bar for the orthogonal contrast Simmental versus (Rendena + Alpine Grey); on Rendena bar for the orthogonal contrast Rendena versus Alpine Grey.
      Figure thumbnail gr3
      Figure 3Breed effect of αS1-CN, αS2-CN, and β-CN expressed as content of milk (g/L) and as proportion of total milk nitrogen (%N) estimated according to a model including herd-date, parity, and lactation stage (base model, in solid color) or also the genotypes of CSN2, CSN3, and BLG (+Gen model, in dashed color). Asterisks (*P < 0.05; **P < 0.01; ***P < 0.001) on the Jersey bar refer to the orthogonal contrast Jersey versus (Holstein + Brown Swiss); on Holstein bar for the orthogonal contrast Holstein versus Brown Swiss; on Simmental bar for the orthogonal contrast Simmental versus (Rendena + Alpine Grey); on Rendena bar for the orthogonal contrast Rendena versus Alpine Grey.
      Figure thumbnail gr4
      Figure 4Breed effect of the total κ-CN and its glycosylated (Glyco-κ-CN) and carbohydrate-free (CF-κ-CN) forms, expressed as content of milk (g/L) and as proportion of total milk nitrogen (%N) estimated according to a model including herd-date, parity, and lactation stage (base model, in solid color) or also the genotypes of CSN2, CSN3, and BLG (+Gen model, in dashed color). Asterisks (*P < 0.05; **P < 0.01; ***P < 0.001) on the Jersey bar refer to the orthogonal contrast Jersey versus (Holstein + Brown Swiss); on Holstein bar for the orthogonal contrast Holstein versus Brown Swiss; on Simmental bar for the orthogonal contrast Simmental versus (Rendena + Alpine Grey); on Rendena bar for the orthogonal contrast Rendena versus Alpine Grey.
      Figure thumbnail gr5
      Figure 5Breed effect of the whey proteins expressed as content of milk (g/L) and as proportion of total milk nitrogen (%N) estimated according to a model including herd-date, parity, and lactation stage (base model, in solid color) or also the genotypes of CSN2, CSN3, and BLG (+Gen model, in dashed color). Asterisks (*P < 0.05; **P < 0.01; ***P < 0.001) on the Jersey bar refer to the orthogonal contrast Jersey versus (Holstein + Brown Swiss); on Holstein bar for the orthogonal contrast Holstein versus Brown Swiss; on Simmental bar for the orthogonal contrast Simmental versus (Rendena + Alpine Grey); on Rendena bar for the orthogonal contrast Rendena versus Alpine Grey.
      Figure thumbnail gr6
      Figure 6Breed effect of the NPN fractions expressed as content of milk (g/L) and as proportion of total milk nitrogen (%N) estimated according to a model including herd-date, parity, and lactation stage (base model, in solid color) or also the genotypes of CSN2, CSN3, and BLG (+Gen model, in dashed color). Asterisks (*P < 0.05; **P < 0.01; ***P < 0.001) on the Jersey bar refer to the orthogonal contrast Jersey versus (Holstein + Brown Swiss); on Holstein bar for the orthogonal contrast Holstein versus Brown Swiss; on Simmental bar for the orthogonal contrast Simmental versus (Rendena + Alpine Grey); on Rendena bar for the orthogonal contrast Rendena versus Alpine Grey.
      In qualitative terms (%N), the greater proportions of NPN compounds and of κ-CN and its glycosylated form were compensated for by smaller proportions of MUN, all the whey proteins, β-CN, and true protein (Figures 2, 3, 4, 5, and 6). After taking into account the effect of the genotypes of CSN2, CSN3, and BLG, the higher proportion of κ-CN in the total N of Jersey milk almost disappeared and was no longer significant (Figure 4), whereas the proportion of αS1-CN increased and became significant (Figure 3). This means that, as far as the latter casein fraction is concerned, most of the differences between the qualitative protein profiles of the Jerseys and the large-framed dairy breeds are only partially explained by differences in the frequencies of the major genetic variants and depend instead principally on the polygenes of the breeds. The significances in the quantitative protein profile (g/L milk) of Jersey milk were all confirmed, with only some minor changes regarding αS2-CN and β-CN (Figure 3).
      Similar results were obtained by
      • Tacoma R.
      • Fields J.
      • Ebenstein D.B.
      • Lam Y.W.
      • Greenwood S.L.
      Characterization of the bovine milk proteome in early-lactation Holstein and Jersey breeds of dairy cows.
      , who found Jersey milk to be superior to Holstein-Friesian milk in all the protein fractions, including the whey proteins.
      • Jensen H.B.
      • Poulsen N.A.
      • Andersen K.K.
      • Hammershøj M.
      • Poulsen H.D.
      • Larsen L.B.
      Distinct composition of bovine milk from Jersey and Holstein-Friesian cows with good, poor, or noncoagulation properties as reflected in protein genetic variants and isoforms.
      and
      • Gustavsson F.
      • Buitenhuis A.J.
      • Johansson M.
      • Bertelsen H.P.
      • Glantz M.
      • Poulsen N.A.
      • Lindmark Månsson H.
      • Stålhammar H.
      • Larsen L.B.
      • Bendixen C.
      • Paulsson M.
      • Andrén A.
      Effects of breed and casein genetic variants on protein profile in milk from Swedish red, Danish Holstein, and Danish Jersey cows.
      reported that Jersey milk also had smaller relative proportions of whey protein and β-CN than Holstein-Friesian milk. However, unlike in this latter work and in our study, in
      • Jensen H.B.
      • Poulsen N.A.
      • Andersen K.K.
      • Hammershøj M.
      • Poulsen H.D.
      • Larsen L.B.
      Distinct composition of bovine milk from Jersey and Holstein-Friesian cows with good, poor, or noncoagulation properties as reflected in protein genetic variants and isoforms.
      the Holstein-Friesian milk had a higher proportion of κ-CN.
      • Cerbulis J.
      • Farrell Jr., H.M.
      Composition of milks of dairy cattle. I. Protein, lactose, and fat contents and distribution of protein fraction.
      found Brown Swiss and Holstein-Friesian milk to have more NPN than Jersey milk.

      Detailed Milk Protein Profiles of Milk from Holstein-Friesian Versus Brown Swiss Cows

      Comparison between the large-framed Holstein-Friesian and Brown Swiss breeds showed the latter to be clearly superior to the former for all protein fractions in quantitative terms (g/L), a result primarily of its higher milk CP content (
      • Mistry V.V.
      • Brouk M.J.
      • Kasperson K.M.
      • Martin E.
      Cheddar cheese from milk of Holstein and Brown Swiss cows.
      ;
      • Cecchinato A.
      • Albera A.
      • Cipolat-Gotet C.
      • Ferragina A.
      • Bittante G.
      Genetic parameters of cheese yield and curd nutrient recovery or whey loss traits predicted using Fourier-transform infrared spectroscopy of samples collected during milk recording on Holstein, Brown Swiss, and Simmental dairy cows.
      ). In terms of protein quality, the Holstein-Friesian milk had a greater proportion of true protein (Figure 2) due to higher proportions of αS1-CN, β-CN, and whey proteins (Figure 3, Figure 5), compensated for by the lower proportions of the αS2-CN (Figure 3) and κ-CN fractions (Figure 4) and the NPN compounds, except MUN (Figure 6). The protein in milk from Holstein cows has also been found to contain a higher proportion of whey proteins and a lower proportion of κ-CN compared with crossbred cows, including Brown Swiss, Montbéliarde, and Swedish Red breeds (
      • Maurmayr A.
      • Pegolo S.
      • Malchiodi F.
      • Bittante G.
      • Cecchinato A.
      Milk protein composition in purebred Holsteins and in first/second-generation crossbred cows from Swedish Red, Montbeliarde and Brown Swiss bulls.
      ).
      It is worth noting that, when the effects of the genotypes of the major genes were included, most of the differences between the 2 larger breeds became weaker and were no longer significant. This means that the differences between the qualitative milk protein profiles of Holstein-Friesian and Brown Swiss depend largely on the major genes, whereas the differences between their quantitative milk protein profiles depend more on the polygenic superiority of the CP content of Brown Swiss milk. Regarding the quantitative protein profile, the superiority of the Brown Swiss cows was also reported by
      • Cerbulis J.
      • Farrell Jr., H.M.
      Composition of milks of dairy cattle. I. Protein, lactose, and fat contents and distribution of protein fraction.
      , who also found Holstein milk to have a lower content and proportion of NPN compounds.

      Detailed Milk Protein Profiles of Milk from Dual-Purpose Breeds

      Very little information is available resulting from comparison of the detailed milk protein profiles of dual-purpose breeds, whether quantitative or qualitative. We compared the large-framed, well-known (at least in Europe), dual-purpose Simmental breed with 2 medium-framed local alpine dual-purpose breeds, Rendena and Alpine Grey. All 3 breeds had similar major milk protein contents (CP, true protein, casein, and β-CN), although compared with the local alpine breeds, Simmental milk had greater contents of αS1-CN and β-LG and lower contents of κ-CN, α-LA, and NPN compounds, except MUN (Figures 2 to 6). Regarding the qualitative profile, Simmental milk had greater proportions of true protein, αS1-CN, and β-LG fractions and smaller proportions of αS2-CN, κ-CN, α-LA, and NPN compounds (Figures 2 to 6). When the frequencies of the genetic variants of the major milk protein genes were taken into account, the differences became even weaker, confirming the important role played by CSN2, CNS3, and BLG in differentiating between the milk produced by different dual-purpose breeds reared in the same herds.
      The protein profile of milk from Simmental cows has been widely studied by
      • Bonfatti V.
      • Di Martino G.
      • Cecchinato A.
      • Degano L.
      • Carnier P.
      Effects of β-κ-casein (CSN2–CSN3) haplotypes, β-lactoglobulin (BLG) genotypes, and detailed protein composition on coagulation properties of individual milk of Simmental cows.
      ,
      • Bonfatti V.
      • Di Martino G.
      • Cecchinato A.
      • Vicario D.
      • Carnier P.
      Effects of β-κ-casein (CSN2–CSN3) haplotypes and β-lactoglobulin (BLG) genotypes on milk production traits and detailed protein composition of individual milk of Simmental cows.
      ). Compared with the population of those studies, our Simmental milk had slightly lower CP and casein contents and a higher whey protein content. Furthermore, the total κ-CN was higher in our study due to the inclusion in our analyses of the glycosylated form.
      Almost no information is available in the scientific literature on the local breeds. In this study, the Rendena milk had lower contents of CP (similar to the Holstein-Friesians) and of almost all the protein fractions (except α-LA and the NPN compounds) compared with the Alpine Grey milk. Qualitatively, the CP of Rendena milk contained proportionally less true protein, κ-CN, and β-LG and more β-CN, α-LA, and NPN compounds (Figures 2 to 6). Again, the differences between the 2 local breeds were attenuated when the genetic variants of the major milk protein genes were taken into account.

      Industry Interest in Detailed Milk Protein Profiles and Genetic Improvement

      The protein profile of milk is very important in relation to the nutritional value of milk and milk-derived dairy products, and particularly so for cheese-making. Previous research (
      • Jõudu I.
      • Henno M.
      • Kaart T.
      • Püssa T.
      • Kärt O.
      The effect of milk protein contents on the rennet coagulation properties of milk from individual dairy cows.
      ;
      • Amalfitano N.
      • Cipolat-Gotet C.
      • Cecchinato A.
      • Malacarne M.
      • Summer A.
      • Bittante G.
      Milk protein fractions strongly affect the patterns of coagulation, curd firming, and syneresis.
      ) clearly indicates that αS1-CN and κ-CN positively affect the coagulation and curd firming processes. Whether expressed quantitatively or qualitatively, αS2-CN always has a negative effect, whereas β-CN has a positive effect when expressed in grams per liter but a negative effect when expressed in %N (negative effect of substitution with other more favorable caseins). However, it should be borne in mind that β-CN is negatively associated with both αS1-CN and κ-CN (
      • Dadousis C.
      • Cipolat-Gotet C.
      • Schiavon S.
      • Bittante G.
      • Cecchinato A.
      Inferring individual cow effects, dairy system effects and feeding effects on latent variables underlying milk protein composition and cheese-making traits in dairy cattle.
      ). The β-LG fraction, too, always had an unfavorable effect on coagulation and curd firming, whereas α-LA had modest effects. An increase in the casein content increases not only the cheese yield, as expected, but also the cheese-making efficiency because it improves the recovery of fat and protein in the cheese and reduces their loss in the whey (
      • Cecchinato A.
      • Albera A.
      • Cipolat-Gotet C.
      • Ferragina A.
      • Bittante G.
      Genetic parameters of cheese yield and curd nutrient recovery or whey loss traits predicted using Fourier-transform infrared spectroscopy of samples collected during milk recording on Holstein, Brown Swiss, and Simmental dairy cows.
      ;
      • Cipolat-Gotet C.
      • Cecchinato A.
      • Malacarne M.
      • Bittante G.
      • Summer A.
      Variations in milk protein fractions affect the efficiency of the cheese-making process.
      ). Again, the different protein fractions of milk exhibited different effects: αS2-CN and β-LG depressed the recovery of both fat and protein, whereas the effects of the other protein fractions were more variable. It should be considered that the effects of αS2-CN are often independent of those of the other caseins and that β-LG is associated unfavorably with protein recovery and favorably with NPN compounds, which in turn are negatively associated with udder health (
      • Dadousis C.
      • Cipolat-Gotet C.
      • Schiavon S.
      • Bittante G.
      • Cecchinato A.
      Inferring individual cow effects, dairy system effects and feeding effects on latent variables underlying milk protein composition and cheese-making traits in dairy cattle.
      ). The overall result is that the milk protein profile is expected to have technical and economic implications for the dairy industry, which should therefore take an interest in monitoring and improving the protein profiles of milk destined for cheese-making.
      This study has shown that detailed protein profiles of milk differ profoundly according to different breeds. The main driver of these differences is not dairy specialization, because the milk from dairy breeds is not, on average, much different from the milk from dual-purpose breeds. In fact, we found large differences within dairy breeds and within dual-purpose breeds. As cows of different breeds were reared and sampled in the same environmental conditions (the same farms, feeding regimens, and sampling dates), and as the data were corrected taking into account the effects of the main individual time-dependent factors of variation (parity and lactation stage), we can assume that the major source of the differences in the milk protein profiles of the different breeds is of genetic origin. Inclusion of the genotypes of the major milk protein genetic variants in one of the models revealed that the relative importance of these genes with respect to the overall effects of breed of cow is very different for different protein fractions. Improvement of the milk protein profile should, therefore, be based on increasing the frequencies of the favorable genetic variants of the major genes of milk proteins and on selection using quantitative genetics and genomic tools to enhance the contribution of polygenes. The relative importance of these 2 tools depends not only on the trait of interest but also on the breed. It is well known that some milk protein fractions have high heritability values in Holstein-Friesian (
      • Kroeker E.M.
      • Ng-Kwai-Hang K.F.
      • Hayes J.F.
      • Moxley J.E.
      Heritabilities of relative percentages of major bovine casein and serum proteins in test-day milk samples.
      ;
      • Schopen G.C.B.
      • Heck J.M.L.
      • Bovenhuis H.
      • Visker M.H.P.W.
      • van Valenberg H.J.F.
      • van Arendonk J.A.M.
      Genetic parameters for major milk proteins in Dutch Holstein-Friesians.
      ), Brown Swiss (
      • Dadousis C.
      • Cipolat-Gotet C.
      • Bittante G.
      • Cecchinato A.
      Inferring genetic parameters on latent variables underlying milk yield and quality, protein composition, curd firmness and cheese-making traits in dairy cattle.
      ), and Simmental (
      • Bonfatti V.
      • Cecchinato A.
      • Gallo L.
      • Blasco A.
      • Carnier P.
      Genetic analysis of detailed milk protein composition and coagulation properties in Simmental cattle.
      ) cows, as do nonprotein fractions (
      • Saavedra-Jiménez L.A.
      • Ramírez-Valverde R.
      • Núñez-Domínguez R.
      • Ruíz-Flores A.
      • García-Muñiz J.G.
      Genetic parameters for nitrogen fractions content in Mexican Brown Swiss cattle milk.
      ). More recently, the genomic bases and associations have also been studied (
      • Bionaz M.
      • Loor J.J.
      Gene networks driving bovine mammary protein synthesis during the lactation cycle.
      ;
      • Dadousis C.
      • Pegolo S.
      • Rosa G.J.M.
      • Bittante G.
      • Cecchinato A.
      Genome-wide association and pathway-based analysis using latent variables related to milk protein composition and cheesemaking traits in dairy cattle.
      ;
      • Pegolo S.
      • Mach N.
      • Ramayo-Caldas Y.
      • Schiavon S.
      • Bittante G.
      • Cecchinato A.
      Integration of GWAS, pathway and network analyses reveals novel mechanistic insights into the synthesis of milk proteins in dairy cows.
      ), leading the way to possible genomic selection in the near future.

      CONCLUSIONS

      In this study we built a detailed profile of 15 milk nitrogenous compounds expressed quantitatively (in g/L) or qualitatively (in % of milk N). There were notable differences between the various breeds of cow for all the N fractions, with the exception of the urea and total casein contents of milk and of the urea and β-CN proportions of total milk N. Inclusion in the model of the genotypes of the major milk protein genes explained a large part of the total variation in the individual caseins and whey protein fractions and in the nonprotein fractions, except urea. That the major genes explain such a large part of the phenotypic variance reduces the phenotypic variance explained mainly by breed and the residual variance in proportions that vary considerably according to the different milk protein traits. The CSN2 gene showed a greater contribution to the phenotypic variance of the contents of αS2-CN and the noncasein traits than that of the contents of β-CN and the other caseins; CSN3 influenced the milk content variability of all the individual casein fractions (especially the κ-CN fractions) as well as α-LA; and the contribution of BLG was notable for the phenotypic variance of the content of all the noncasein fractions, except α-LA and urea. We were able to compare the effect of breed with and without inclusion in the models of the genotypes of the major protein genetic variants with respect to both the qualitative and quantitative milk protein profiles. This means that these 3 major genes explain (1) a notable portion of the total phenotypic variance in the detailed milk protein profile, (2) the greatest part of the differences between the detailed protein profiles of milk from different breeds, and (3) an important part of the genetic variation among different cows within the same breed. This information will provide the dairy industry with a better means of evaluating milk from different breeds of cow, especially where it is destined for cheese-making. It will also be of use in developing the best strategies for genetically improving different breeds for different milk protein traits.

      ACKNOWLEDGMENTS

      The authors thank the Autonomous Province of Trento (Italy), which provided the funds for the Cowplus project. The authors confirm that there were no conflicts of interest.

      REFERENCES

        • Aguilar M.
        • Hanigan M.D.
        • Tucker H.A.
        • Jones B.L.
        • Garbade S.K.
        • McGilliard M.L.
        • Stallings C.C.
        • Knowlton K.F.
        • James R.E.
        Cow and herd variation in milk urea nitrogen concentrations in lactating dairy cattle.
        J. Dairy Sci. 2012; 95 (23040023): 7261-7268
        • Ali A.K.A.
        • Shook G.E.
        An optimum transformation for somatic cell concentration in milk.
        J. Dairy Sci. 1980; 63: 487-490
        • Amalfitano N.
        • Cipolat-Gotet C.
        • Cecchinato A.
        • Malacarne M.
        • Summer A.
        • Bittante G.
        Milk protein fractions strongly affect the patterns of coagulation, curd firming, and syneresis.
        J. Dairy Sci. 2019; 102 (30772026): 2903-2917
        • Barłowska J.
        • Szwajkowska M.
        • Litwińczuk Z.
        • Kròl J.
        Nutritional value and technological suitability of milk from various animal species used for dairy production.
        Compr. Rev. Food Sci. Food Saf. 2011; 10: 291-302
        • Bastin C.
        • Laloux L.
        • Gillon A.
        • Miglior F.
        • Soyeurt H.
        • Hammami H.
        • Bertozzi C.
        • Gengler N.
        Modeling milk urea of Walloon dairy cows in management perspectives.
        J. Dairy Sci. 2009; 92 (19528631): 3529-3540
        • Bionaz M.
        • Loor J.J.
        Gene networks driving bovine mammary protein synthesis during the lactation cycle.
        Bioinform. Biol. Insights. 2011; 5 (21698073): 83-98
        • Bittante G.
        • Penasa M.
        • Cecchinato A.
        Invited review: Genetics and modeling of milk coagulation properties.
        J. Dairy Sci. 2012; 95 (23021752): 6843-6870
        • Bland J.H.
        • Grandison A.S.
        • Fagan C.C.
        Effect of blending Jersey and Holstein-Friesian milk on Cheddar cheese processing, composition, and quality.
        J. Dairy Sci. 2015; 98 (25465548): 1-8
        • Bobbo T.
        • Fiore E.
        • Gianesella M.
        • Morgante M.
        • Gallo L.
        • Ruegg P.L.
        • Bittante G.
        • Cecchinato A.
        Variation in blood serum proteins and association with somatic cell count in dairy cattle from multi-breed herds.
        Animal. 2017; 11 (28560948): 2309-2319
        • Bobbo T.
        • Ruegg P.L.
        • Stocco G.
        • Fiore E.
        • Gianesella M.
        • Morgante M.
        • Pasotto D.
        • Bittante G.
        • Cecchinato A.
        Associations between pathogen-specific cases of subclinical mastitis and milk yield, quality, protein composition, and cheese-making traits in dairy cows.
        J. Dairy Sci. 2017; 100 (28365113): 4868-4883
        • Bobe G.
        • Beitz D.C.
        • Freeman A.E.
        • Lindberg G.L.
        Effect of milk protein genotypes on milk protein composition and its genetic parameter estimates.
        J. Dairy Sci. 1999; 82 (10629828): 2797-2804
        • Bonfatti V.
        • Cecchinato A.
        • Gallo L.
        • Blasco A.
        • Carnier P.
        Genetic analysis of detailed milk protein composition and coagulation properties in Simmental cattle.
        J. Dairy Sci. 2011; 94 (21943768): 5183-5193
        • Bonfatti V.
        • Di Martino G.
        • Cecchinato A.
        • Degano L.
        • Carnier P.
        Effects of β-κ-casein (CSN2–CSN3) haplotypes, β-lactoglobulin (BLG) genotypes, and detailed protein composition on coagulation properties of individual milk of Simmental cows.
        J. Dairy Sci. 2010; 93 (20655451): 3809-3817
        • Bonfatti V.
        • Di Martino G.
        • Cecchinato A.
        • Vicario D.
        • Carnier P.
        Effects of β-κ-casein (CSN2–CSN3) haplotypes and β-lactoglobulin (BLG) genotypes on milk production traits and detailed protein composition of individual milk of Simmental cows.
        J. Dairy Sci. 2010; 93 (20655450): 3797-3808
        • Caroli A.M.
        • Chessa S.
        • Erhardt G.J.
        Invited review: Milk protein polymorphisms in cattle: Effect on animal breeding and human nutrition.
        J. Dairy Sci. 2009; 92 (19841193): 5335-5352
        • Cecchinato A.
        • Albera A.
        • Cipolat-Gotet C.
        • Ferragina A.
        • Bittante G.
        Genetic parameters of cheese yield and curd nutrient recovery or whey loss traits predicted using Fourier-transform infrared spectroscopy of samples collected during milk recording on Holstein, Brown Swiss, and Simmental dairy cows.
        J. Dairy Sci. 2015; 98 (25958274): 4914-4927
        • Cerbulis J.
        • Farrell Jr., H.M.
        Composition of milks of dairy cattle. I. Protein, lactose, and fat contents and distribution of protein fraction.
        J. Dairy Sci. 1975; 58 (1141480): 817-827
        • Cipolat-Gotet C.
        • Cecchinato A.
        • Malacarne M.
        • Bittante G.
        • Summer A.
        Variations in milk protein fractions affect the efficiency of the cheese-making process.
        J. Dairy Sci. 2018; 101 (30122415): 8788-8804
        • Dadousis C.
        • Cipolat-Gotet C.
        • Bittante G.
        • Cecchinato A.
        Inferring genetic parameters on latent variables underlying milk yield and quality, protein composition, curd firmness and cheese-making traits in dairy cattle.
        Animal. 2018; 12 (28712368): 224-231
        • Dadousis C.
        • Cipolat-Gotet C.
        • Schiavon S.
        • Bittante G.
        • Cecchinato A.
        Inferring individual cow effects, dairy system effects and feeding effects on latent variables underlying milk protein composition and cheese-making traits in dairy cattle.
        J. Dairy Res. 2018; 85 (29125094): 87-97
        • Dadousis C.
        • Pegolo S.
        • Rosa G.J.M.
        • Bittante G.
        • Cecchinato A.
        Genome-wide association and pathway-based analysis using latent variables related to milk protein composition and cheesemaking traits in dairy cattle.
        J. Dairy Sci. 2017; 100 (28843680): 9085-9102
        • Gustavsson F.
        • Buitenhuis A.J.
        • Johansson M.
        • Bertelsen H.P.
        • Glantz M.
        • Poulsen N.A.
        • Lindmark Månsson H.
        • Stålhammar H.
        • Larsen L.B.
        • Bendixen C.
        • Paulsson M.
        • Andrén A.
        Effects of breed and casein genetic variants on protein profile in milk from Swedish red, Danish Holstein, and Danish Jersey cows.
        J. Dairy Sci. 2014; 97 (24704225): 3866-3877
        • Hallén E.
        • Wedholm A.
        • Andrén A.
        • Lundén A.
        Effect of β-casein, κ-casein and β-lactoglobulin genotypes on concentration of milk protein variants.
        J. Anim. Breed. Genet. 2008; 125 (18363977): 119-129
        • Haug A.
        • Høstmark A.T.
        • Harstad O.M.
        Bovine milk in human nutrition—A review.
        Lipids Health Dis. 2007; 6 (17894873): 25
        • Heck J.M.L.
        • Olieman C.
        • Schennink A.
        • van Valenberg H.J.F.
        • Visker M.H.P.W.
        • Meuldijk R.C.R.
        • van Hooijdonk A.C.M.
        Estimation of variation in concentration, phosphorylation and genetic polymorphism of milk proteins using capillary zone electrophoresis.
        Int. Dairy J. 2008; 18: 548-555
        • Heck J.M.L.
        • Schennink A.
        • Van Valenberg H.J.F.
        • Bovenhuis H.
        • Visker M.H.P.W.
        • Van Arendonk J.A.M.
        • Van Hooijdonk A.C.M.
        Effects of milk protein variants on the protein composition of bovine milk.
        J. Dairy Sci. 2009; 92 (19233813): 1192-1202
        • Hoffman P.C.
        • Esser N.M.
        • Bauman L.M.
        • Denzine S.L.
        • Engstrom M.
        • Chester-Jones H.
        Short communication: Effect of dietary protein on growth and nitrogen balance of Holstein heifers.
        J. Dairy Sci. 2001; 84 (11352161): 843-847
        • Jensen H.B.
        • Poulsen N.A.
        • Andersen K.K.
        • Hammershøj M.
        • Poulsen H.D.
        • Larsen L.B.
        Distinct composition of bovine milk from Jersey and Holstein-Friesian cows with good, poor, or noncoagulation properties as reflected in protein genetic variants and isoforms.
        J. Dairy Sci. 2012; 95 (23040012): 6905-6917
        • Jõudu I.
        • Henno M.
        • Kaart T.
        • Püssa T.
        • Kärt O.
        The effect of milk protein contents on the rennet coagulation properties of milk from individual dairy cows.
        Int. Dairy J. 2008; 18: 964-967
        • Korhonen H.
        • Pihlanto A.
        Bioactive peptides: Production and functionality.
        Int. Dairy J. 2006; 16: 945-960
        • Kroeker E.M.
        • Ng-Kwai-Hang K.F.
        • Hayes J.F.
        • Moxley J.E.
        Heritabilities of relative percentages of major bovine casein and serum proteins in test-day milk samples.
        J. Dairy Sci. 1985; 68: 1346-1348
        • Law B.A.
        • Tamime A.Y.
        Technology of Cheese-Making.
        2nd ed. Wiley-Blackwell, Chichester, UK2010
        • Law R.A.
        • Young F.J.
        • Patterson D.C.
        • Kilpatrick D.J.
        • Wylie A.R.G.
        • Mayne C.S.
        Effect of dietary protein content on animal production and blood metabolites of dairy cows during lactation.
        J. Dairy Sci. 2009; 92 (19233794): 1001-1012
        • Martin P.
        • Szymanowska M.
        • Zwierzchowski L.
        • Leroux C.
        The impact of genetic polymorphisms on the protein composition of ruminant milks.
        Reprod. Nutr. Dev. 2002; 42 (12537255): 433-459
        • Maurmayr A.
        • Cecchinato A.
        • Grigoletto L.
        • Bittante G.
        Detection and quantification of α S1-, α S2-, β-, κ-casein, α-lactalbumin, β-lactoglobulin and lactoferrin in bovine milk by reverse-phase high-performance liquid chromatography.
        Agriculturae Conspectus Scientificus. 2013; 78: 201-205
        • Maurmayr A.
        • Pegolo S.
        • Malchiodi F.
        • Bittante G.
        • Cecchinato A.
        Milk protein composition in purebred Holsteins and in first/second-generation crossbred cows from Swedish Red, Montbeliarde and Brown Swiss bulls.
        Animal. 2018; 12 (29307328): 2214-2220
        • Mistry V.V.
        • Brouk M.J.
        • Kasperson K.M.
        • Martin E.
        Cheddar cheese from milk of Holstein and Brown Swiss cows.
        Milchwissenschft. 2002; 57: 19-23
        • Ng-Kwai-Hang K.F.
        • Hayes J.F.
        • Moxley J.E.
        • Monardes H.G.
        Variation in milk protein concentrations associated with genetic polymorphism and environmental factors.
        J. Dairy Sci. 1987; 70 (3584600): 563-570
        • Pegolo S.
        • Mach N.
        • Ramayo-Caldas Y.
        • Schiavon S.
        • Bittante G.
        • Cecchinato A.
        Integration of GWAS, pathway and network analyses reveals novel mechanistic insights into the synthesis of milk proteins in dairy cows.
        Sci. Rep. 2018; 8 (29330500): 566
        • Poulsen N.A.
        • Jensen H.B.
        • Larsen L.B.
        Factors influencing degree of glycosylation and phosphorylation of caseins in individual cow milk samples.
        J. Dairy Sci. 2016; 99 (26995120): 3325-3333
        • Raggio G.
        • Lobley G.E.
        • Berthiaume R.
        • Pellerin D.
        • Allard G.
        • Dubreuil P.
        • Lapierre H.
        Effect of protein supply on hepatic synthesis of plasma and constitutive proteins in lactating dairy cows.
        J. Dairy Sci. 2007; 90 (17183103): 352-359
        • Robitaille G.
        • Britten M.
        • Morisset J.
        • Petitclerc D.
        Quantitative analysis of β-lactoglobulin A and B genetic variants in milk of cows β-lactoglobulin AB throughout lactation.
        J. Dairy Res. 2002; 69 (12463701): 651-654
        • Rowlands G.J.
        • Little W.
        • Kitchenham B.A.
        Relationships between blood composition and fertility in dairy cows—A field study.
        J. Dairy Res. 1977; 44 (856897): 1-7
        • Saavedra-Jiménez L.A.
        • Ramírez-Valverde R.
        • Núñez-Domínguez R.
        • Ruíz-Flores A.
        • García-Muñiz J.G.
        Genetic parameters for nitrogen fractions content in Mexican Brown Swiss cattle milk.
        Trop. Anim. Health Prod. 2019; 51 (31140119): 2235-2241
        • Schiavon S.
        • Cesaro G.
        • Tagliapietra F.
        • Gallo L.
        • Bittante G.
        Influence of N shortage and conjugated linoleic acid supplementation on some productive, digestive, and metabolic parameters of lactating cows.
        Anim. Feed Sci. Technol. 2015; 208: 86-97
        • Schiavon S.
        • Sturaro E.
        • Tagliapietra F.
        • Ramanzin M.
        • Bittante G.
        Nitrogen and phosphorus excretion on mountain farms of different dairy systems.
        Agric. Syst. 2019; 168: 36-47
        • Schopen G.C.B.
        • Heck J.M.L.
        • Bovenhuis H.
        • Visker M.H.P.W.
        • van Valenberg H.J.F.
        • van Arendonk J.A.M.
        Genetic parameters for major milk proteins in Dutch Holstein-Friesians.
        J. Dairy Sci. 2009; 92 (19233812): 1182-1191
        • Stocco G.
        • Cipolat-Gotet C.
        • Bobbo T.
        • Cecchinato A.
        • Bittante G.
        Breed of cow and herd productivity affect milk composition and modeling of coagulation, curd firming, and syneresis.
        J. Dairy Sci. 2017; 100 (27837976): 129-145
        • Stocco G.
        • Cipolat-Gotet C.
        • Gasparotto V.
        • Cecchinato A.
        • Bittante G.
        Breed of cow and herd productivity affect milk nutrient recovery in curd, and cheese yield, efficiency and daily production.
        Animal. 2018; 12 (28712377): 434-444
        • Tacoma R.
        • Fields J.
        • Ebenstein D.B.
        • Lam Y.W.
        • Greenwood S.L.
        Characterization of the bovine milk proteome in early-lactation Holstein and Jersey breeds of dairy cows.
        J. Proteomics. 2016; 130 (26391770): 200-210
        • Wittenburg D.
        • Melzer N.
        • Willmitzer L.
        • Lisec J.
        • Kesting U.
        • Reinsch N.
        • Repsilber D.
        Milk metabolites and their genetic variability.
        J. Dairy Sci. 2013; 96 (23403187): 2557-2569